- REP News
- Posts
- Rep AI: Product News November 2023
Rep AI: Product News November 2023
Multilingual Is Here! | Localized Currency | Advanced Search Capabilities | And much more...
Webinar Announcement
Greetings everyone! Fall is here and BFCM (Black Friday, CyberMonday) are upon us. And we have a ton of updates to share, but before we do, we’re having a webinar next Tuesday, called:
Beyond BFCM: How Real DTC Brands Are Skyrocketing Sales With AI
We’ll cover:
The 95% Problem every merchant faces
The Conversational Effect and its impact on sales
The arrival of Relational Commerce
You can register here. And no worries if you can’t make it. There will be a recording available afterward.
Ok, let’s get into some updates.
Rep Updates
Localization & Multilingual Update
Localized Currencies
Shopping online should feel as local as buying from your neighborhood store. That's why Rep now automatically adjusts prices to your shopper's local currency, making their shopping experience seamless and personalized.
Like so:
And that’s only the beginning…
Multilingual
In our global marketplace, language should never be a barrier. With Rep's new multilingual abilities, your store speaks the language of your customers, creating a welcoming and inclusive shopping experience for everyone.
Check it out:
¿Hablas Español?
Rep will now respond in your, the merchant’s, chosen language. It can also chat in the language of the shopper’s location. You decide.
¡Feliz compra!
Improved Conversational Search
This will probably be the most technical update for this month. If you’re less inclined to understand the technological nuances, here’s the TL;DR quick summary…
Summary
As you may or may not know, on-site search can make or break eCommerce success. Did you know:
69% of consumers frequently encounter irrelevant search results while shopping, despite brands claiming their on-site search results are relevant (had to bold that part).
76% of consumers are more likely to make a purchase from sites with accurate search and useful filtering options.
It’s an area that we believe that Rep will solve in the most pleasantly surprising ways, and we’ve just taken a big leap in making that happen.
Rep now understands what customers are searching for on a much deeper level and can sift through everything your store has to offer to find the best matches—no matter how large your store gets.
Put another way, expect an improved search & discovery experience that generates the most relevant responses based on two advanced technologies:
Vector Embedding
Elasticsearch
So? So now your AI concierge can now handle queries like:
For Search: "I'm looking for a black dress for a cocktail party"
For Discovery: "Hey, I'm looking for something for a cocktail party"
Since your concierge knows your catalog inside-out, it can instantly bring up everything that’s contextually relevant to every shopper’s query.
This capability brings us one step closer to the full-service concierge experience we envision and fulfilling the promise of AI for your business.
This is now LIVE, with much fine-tuning and improvements to happen along the way to unleash the full power of Conversational Search.
REP knows your brand. It knows your catalog. Pretty soon, it will know every customer.
Ok, here’s a more technical explanation of what’s happening behind the scenes and a quick breakdown of how this works…
More Technical Explainer Begins Here…
So, our dev team has implemented two very advanced capabilities to enable a better search & discovery experience.
Vector Embedding
Elasticsearch
I’ll try to explain them to you.
Vector Embedding
In the digital world, vector embedding basically turns words, sentences, or documents into a list of numbers (a vector) that represents their meaning. The closer two vectors are in any given numerical space, the more similar their meaning—kinda like a semantic relationship (like water + wet).
Elasticsearch:
Think of Elasticsearch as a super-efficient librarian. It searches through vast amounts of data (like books in a library) to find exactly what you're looking for.
It's designed to handle complex searches and give you results in real-time, which is perfect for online stores like yours where customers expect quick responses.
Here’s How They Work Together:
Search & Discover Experience: When a customer asks a question, the AI needs to understand what they're looking for and then find the best answers from all the searchable content in your store, like product descriptions, FAQs, etc.
Vector Embedding in Action: As soon as a customer types in a query, vector embedding converts this query into a numerical vector that represents the meaning of their question.
The Hand-off to Elasticsearch: With the help of vector embedding, Elasticsearch can now search through all the documents (product pages, collection pages, files, FAQs) not just by keywords but by the meaning inherent in the vectors. Your AI ‘librarian’ can understand the general topic area you're interested in, not just the exact words you used.
Feeding ChatGPT: The best matches found by Elasticsearch are then fed to ChatGPT as "knowledge." This means that ChatGPT doesn't just use pre-programmed responses; it uses the most relevant, up-to-date information from your store to respond to customers (since Rep’s AI is always scanning your catalog).
Scale and Size: Catalog size won’t matter anymore. Whether you have a hundred products or a hundred thousand, vector embedding and Elasticsearch should be able to handle it. As your store grows, you can be confident that the search & discovery experience will remain fast and accurate.
Did you get all that? Great, let’s move on to…
Conversations Screen
Right now, you use a mix of scripted flows with AI-generated answers for responses. Now, when you view recorded conversations, you will see a ChatGPT badge for every response that was generated by AI.
Like so:
Brand Voice Has A New Home
If you need to adjust your tone of voice, it’s now located under AI Training, right where it belongs.
A Better Mobile Experience
We’re working on our mobile experience, making it even more responsive. More updates are on the way.
Product Comparisons
Choosing between products can be overwhelming, especially when the differences aren't clear-cut.
To tackle this, we've introduced Product Comparisons.
Now, shoppers can ask direct questions like, "What's the difference between Product A and Product B?"
It’s not just a side-by-side feature comparison; The AI should understand the nuances of each product to better guide shoppers to the choice that best fits their needs. This means less confusion, more confident decisions, and a smoother path to purchase.
Nothing beats shopper confidence.
Btw, we mentioned this in the last newsletter, but wanted to give you a little more context.
Spotlight: Snow Teeth Whitening
We just published a pretty incredible case study. If you ever want to know what’s possible with AI, please read this one.
I’ll also be featuring Snow’s story in the webinar.
Enjoy,
The Rep Team